Bot2.0: The Memex and Social Learning

1. Bot2.0: The Memex and Social Learning ASIS&T, Columbus, OH October 27, 2008 Jacob Kramer-Duffield, Ph.D. Student Jane Greenberg, Associative Professor and Director, SILS Metadata Research Center School of Information and Library Science, University of North Carolina at Chapel Hill

2. Presentation Overview • In this presentation I will introduce the broad outlines of the Bot2.0 Project • I will present our initial findings from a pre BotCamp survey regarding students' metadata literacy • I will then present and discuss the results of an experiment on structured vs. unstructured knowledge • Finally, I will present broad conclusions and suggest areas for future research

3. Overview • BOT 2.0 is an NSF CCLI- funded project investigating approaches to recruitment, retention, and increased collaborative learning in science education. • BOT 2.0 combines an invitational hands-on weekend program called BotCamp with metadata approaches and Web 2.0 cloud memex technology. • Our research involved student metadata literacy, tagging vs. structured metadata for plant identification, and metadata usage in the context of plant identification exercises • The motivation for the project was to begin utilizing innovative technological and instructional approaches in pursuit of a less-intimidating introductory approach to botany instruction, with the goal of attracting and retaining non-traditional students to the field

4. “Cloud Memex” Technology Our conception of a memex is not a centralized black box but rather a flexible, distributed aggregation of the best services from the public Web 2.0 “cloud.” “Cloud Memex” Technology Our conception of a memex is not a centralized black box but rather a flexible, distributed aggregation of the best services from the public Web 2.0 “cloud.”

5. BotCamp I Participants • BotCamp targeted undergraduate minorities and historically disadvantaged students from a range of institutions in the Piedmont region of North Carolina.

6. Metadata Literacy Survey • Before they arrived we did an online assessment of their baseline knowledge of technology, metadata, and their backgrounds

7. “Metadata Literacy” Survey Data “Have you added labels/tags/descriptions to someone else’s (a person or group’s) digital photographs on the web?” Yes 10 No 4 “Metadata Literacy” Survey Data

8. “Metadata Literacy” Survey Data “Have you heard the word metadata before? If yes, please describe metadata.” Yes 2 No 10 “...words that describe data,” “data that helps search engines find what [you are] looking for”

9. “Metadata Literacy” Survey Data “Have you heard the word Folksonomy before?” Yes 0 No 14

10. Metadata Literacy – Initial Findings • Students are not broadly familiar with metadata terminology, but are comfortable in and enthusiastic towards the use of tagging in practice. • The informal tagging approach suggested a lower barrier toward learning more formal classification systems • Students recognize the importance of standardized terminology

11. Metadata Experiment • We performed an experiment testing students' plant description and identification abilities (and confidence in those abilities)‏ • We were interested in exploring differences that might emerge in these beginning students flowing from varying introductory approaches to understanding knowledge and knowledge organization

12. Metadata Experiment • The plant ID metadata experiment assessed students’ understanding and use of structured metadata, taxonomies, and unstructured tagging-based approaches for learning about botanical science. • Students were split into two groups for a short plant-identification exercise. One group was given a short lesson on tagging, folksonomies and their potential uses in plant identification and learning a new domain. The other group was instructed on taxonomy and its role in botany and plant identification. • In the course of the assessment, students were asked to describe the characters, character states and give as many names as they wished for four plant samples.

13. Metadata Experiment

14. Metadata Experiment • Differences were attributable to differing instructions for the same exercise, but also reflect the general difference of an unstructured, tagging-based and more-exhaustive process of description versus a more-precise process of description. • Most students did not offer confidence ratings for their identifications, but of those that did all who were very confident of their identifications were in the taxonomy condition, and all who assessed in the tagging condition were not confident of their identification.

15. Metadata Experiment • All who responded to the question “I would like to find out more about botany in the future” (seven of the 14 across both conditions) said that they strongly agreed with the statement—a difficult exercise did not deter them. • In direct feedback and later in focus groups, it became clear that the difference between characters and character states and their respective definitions was confusing for all

16. Moving Novices Towards the Expert's Semantic Space • Data suggest that a tagging-based approach can function as a learning bridge, providing a cognitive sign-post which can then be re-directed towards more structured taxonomies at later stages and with the assistance of instruction and expert input. • e.g.,"rounded sections" is a reasonable enough description of many oak leaves—when informed that the feature can be thought of as two botanical character states, "lobed leaves" and "rounded tips,” students can then more easily associate their initial, common-sense classification with a more botanically precise one.

17. Moving Novices Towards the Expert's Semantic Space • Botany in particular is fertile ground for this approach, as field botanists commonly utilize an approach to identification of unknown species ery similar to a web-based free-tagging approach to categorization. • Further integrating these established uses and the other affordances of social technologies will allow for greater exploration of the unstructured-to-structured approach to new knowledge acquisition.

18. Metadata Continuum of Learning Science • From unstructured to structured? • Students were positive in reaction to use of unstructured vocabularies: “Being able to see what other people in my group were calling, I mean labelling, plants was really helpful in the learning process.  It made the lesson feel more collaborative and we could figure out where we were in relation to others that were studying the same plants.”

19. Metadata Continuum of Learning Science • Assumptions made about previous social software experience were mixed in support – nearly all students had Facebook accounts and used Ning easily but few had used Flickr, though their experience there was positive • “I really hadn’t used the tagging function in Flickr before, but doing tagging at BotCamp… I’m hooked.  It makes the photos have more meaning.”

20. Future Research Bot 2.0 ++ • In 2009 we will conduct round two of BOT 2.0 to generate additional data. Building on the successes and lessons learned from the first iteration of BotCamp, we will fine-tune the curriculum, pedagogy and implementation of technologies in support of continued innovations. • My larger research interests revolve around the basic issues of identity underlying students' interfaces with both technology and educational norms, which I will explore further in this ongoing research

21. Acknowledgements and Contact Information Bot 2.0 is a collaboration supported by National Science Foundation Grant # 0737466 involving the North Carolina Botanical Garden, the UNC SILS Metadata Research center, and the Renaissance Computing Institute (RENCI) at UNC. Michael Shoffner, Evelyn Daniel, Stephen Seiberling, Alan Weakley, David Woodbury and Hill Taylor contributed to this work. Web page: http://bot2.org Contacts: Jane Greenberg: janeg@email.unc.eduJacob Kramer-Duffield: jkd@email.unc.edu